Methodology
AI-Assisted Assessments
Where AI may help, where it must not, and how every AI output is tied to evidence and confidence.
Prototype Notice:This prototype currently uses mock and illustrative data. Source records, update schedules, confidence scores and methodology examples are placeholders for the future production system.
Causation rule: Market movement after an event does not by itself prove that the event caused the movement.
AI may be used to
- Summarize verified facts.
- Compare countries on documented indicators.
- Explain indicator changes with cited evidence.
- Identify possible relationships, flagged as hypotheses.
- Generate structured event assessments.
- Draft scenario descriptions with assumptions.
- Highlight uncertainty and gaps.
- Detect inconsistencies for human review.
AI must not
- Invent values or sources.
- Replace official data.
- Present speculation as fact.
- Claim causation without supporting evidence.
- Issue official credit ratings.
- Provide personal investment, legal, tax or medical advice.
- Hide uncertainty.
- Present scenarios as guaranteed outcomes.
AI Assessment Template
Reusable Component
Assessment
Illustrative — interpretive summary of a verified fact.
Supporting Evidence
- Verified fact A (source record placeholder)
- Verified fact B (source record placeholder)
Contradicting Evidence
- Counter-signal X (source record placeholder)
Confidence
Moderate (62/100)
Main Uncertainty
Limited cross-source confirmation in the prototype dataset.
Generated At
2026-06-22T09:14:00Z
Reviewer
Reviewer placeholder
Method
Method placeholder — prototype
Version
ai-method 0.1.0
Related: Confidence · Intelligence Events